Simple Models of Land Hydroclimate Under Warming

Abstract

Land surfaces can dry out, while ocean surfaces provide a steady supply of water. Because of this, climate dynamics are fundamentally different over land versus ocean, but many basic questions about land climate remain unanswered: for example, whether soils will get wetter or drier as temperatures rise, and how water limitation affects mean precipitation under warming ("hydrological sensitivity"). Here we explore these questions with model hierarchies that span a range of complexity, including simple physical models, idealized and full-complexity climate simulations, and cloud-permitting simulations over idealized land surfaces in radiative-convective equilibrium (RCE). We focus especially on simple models, which anchor observed and simulated impacts to conceptual understanding. For soil drying, full-complexity models diverge on the sign of expected changes, despite the common view that soils dry under warming. Using a new physical theory for surface soil moisture, we diagnose soil moisture trends at large scales and explain spatial variability in the current climate (Chapter 2). The theory is radically simpler than published alternatives, dependent only on precipitation and surface net radiation with no free parameters, and no dependence on vapor pressure deficit or plant responses to CO2. We then analyze soil moisture trends in climate models, and show that soils do not dry in the quasi-global mean if integrated deeper than the uppermost soil layers (Chapter 3). A second theory explains patchy regional trends, and predicts drying in regions where the net lateral land and atmosphere water inflow is below a fixed threshold. This condition holds in RCE by definition, and in specific cases across a hierarchy of climate models. Finally, we develop a broader conceptual framework based on this predictive drying model to establish how soil moisture modulates terrestrial hydrological sensitivity via surface conductance (Chapter 4). Together, these basic physical models underscore that warming can have surprising effects when water supply is limited, and does not necessarily imply drying over land.